Part 5: Results and Conclusion
Introduction
What the problem is:
- scope
- next
Summary of Results
| DC minutes | Philly minutes | NYC minutes | DC miles | Philly miles | NYC miles | |
|---|---|---|---|---|---|---|
| Min | 0.00000 | 0.00000 | 0.0000 | 0.000000 | 0.000000 | 0.000000 |
| 1st Qu | 55.30000 | 80.31667 | 83.1500 | 4.250792 | 6.364693 | 11.294021 |
| Median | 72.03333 | 107.78333 | 111.7500 | 6.877634 | 10.163438 | 18.151150 |
| Mean | 72.52182 | 116.91172 | 117.0508 | 7.119155 | 10.623455 | 19.113244 |
| 3rd Qu | 89.18333 | 138.43333 | 141.0000 | 9.775393 | 14.337647 | 26.475954 |
| Max | 163.83333 | 516.33333 | 1316.0500 | 19.387365 | 41.823172 | 68.265572 |
| St. Dev. | 24.61760 | 61.82033 | 79.8526 | 3.620629 | 5.549423 | 9.939558 |
What to say here:
- dc looking good
- least amount of skew (look at that close mean and median)
- lowest everything
- learned later on that this is partially due to DC
- NY and Philly are pretty close
- ew, look at those max values (we’ll look at that soon)
- the difference in sd between minutes and miles is interesting
- maybe google how to look into that
histograms

Clearly some issues with NY and philadelphia:
- as suspected, DC is beautiful
- the skew of Philly and NYC is surprisingly bad
scatterplots

Looking at these summary statistics and histogram, I would like to put forward an axiom:
Any trip over three hours, in this situation, can be defined as “Shit Luck”
good points are in light blue, Bad points are in dark blue, truly evil points in black:
- let’s split this shit up
- normal < 180
- 180 < bad < 300
- evil > 300

Percent of Bad Trips
| DC | PH | NY | |
|---|---|---|---|
| 3 hour trip | 0.00 | 8.12 | 5.98 |
| 2 hour trip | 2.96 | 38.85 | 42.65 |
Here I will talk about
this shit cause
Im dope as fuck it is the percent of bad trips in each city
Here are the scatterplots of only trips less than 3 hours shows a pretty nice casual relationship discuss each DC being a tiny city is hugely helpful

The “bad luck” trips
Both New York City and Philadelphia have some weird clusters of points evident from their scatter plot. Let’s look at the trips that take more than 400 minutes.
| start_latitude | Freq | end_latitude | Freq | start_latitude | Freq | end_latitude | Freq |
|---|---|---|---|---|---|---|---|
| 39.8888 | 288 | 39.8888 | 94 | 40.5404671945649 | 259 | NA | NA |
| 39.9010033 | 93 | 39.8911694 | 18 | NA | NA | NA | |
| 40.0535197 | 33 | 39.9755981 | 3 | NA | NA | NA | |
| 39.8840912 | 2 | 40.0324002 | 3 | NA | NA | NA | |
| 39.9413149 | 2 | 40.0514994 | 3 | NA | NA | NA | |
| 39.9541915 | 2 | 40.0526994 | 3 | NA | NA | NA |
Here:
- several points in Philly
- Only 1 in NYC
- WTF is up with these points?
The EVIL NY POINT
Here is:
- The routes
- That one point everything is going towards

Philadelphia
Here is:
- The routee
- there are several points that everything comes/goes to
- see leaflet and rewrite

What’s happening here:
- I have take a sample of 50 routes from the top 10% of trips, bottom 10% of cities, and middle 50
- plotted those routes
Sample of Short, Medium, and Long Trips

Neighborhoods individual
okay so app being in the rivers or oceans can be both really negative and realy positive
One of the coolest things about NYC is the spread of the best locations to start. DC and Philly, they’re very bunched up.